Methanol to Olefins (MTO): From Fundamentals toCommercializationPeng Tian,† Yingxu Wei,† Mao Ye,† and Zhongmin Liu*,†,‡

†National Engineering Laboratory for Methanol to Olefins, Dalian National Laboratory for Clean Energy, Dalian Institute ofChemical Physics, Chinese Academy of Sciences, Dalian 116023, China‡State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China

*S Supporting Information

ABSTRACT: The methanol-to-olefins (MTO) reaction is aninteresting and important reaction for both fundamental researchand industrial application. The Dalian Institute of ChemicalPhysics (DICP) has developed a MTO technology that led to thesuccessful construction and operation of the world’s first coal toolefin plant in 2010. This historical perspective gives a briefsummary on the key issues for the process development, includingstudies on the reaction mechanism, molecular sieve synthesis andcrystallization mechanism, catalyst and its manufacturing scale-up,reactor selection and reactor scale-up, process demonstration, andcommercialization. Further challenges on the fundamental researchand the directions for future catalyst improvement are alsosuggested.

KEYWORDS: methanol to olefins, DMTO, reaction mechanism, SAPO molecular sieve, scale-up, fluidized bed

1. INTRODUCTION

The methanol-to-olefins (MTO) reaction is one of the mostimportant reactions in C1 chemistry, which provides a chancefor producing basic petrochemicals from nonoil resources suchas coal and natural gas. As olefin-based petrochemicals andrelevant downstream processes have been well developed formany years, MTO is believed to be a linkage between coal ornatural gas chemical industry and modern petrochemicalindustry. Many institutions and companies have put greateffort to the research of MTO reaction since it was firstproposed by Mobil Corporation in 1977,1 and significantprogress has been achieved with respect to the reactionprinciple,2−10 catalyst synthesis,11−15 and process research anddevelopment (R&D).16−19

Methanol, which is very sensitive to a catalyst due to its highactivity, could be catalyzed by acidic zeolites to formhydrocarbons. The reactant molecule is small and simple, butthe reaction has been demonstrated to be very complicatedwith a large variety of products over different zeolite catalysts.The successful development of a commercially applicable MTOtechnology needs to solve many scientific and technicalproblems: (1) to establish a selectivity control principle forthis complex reaction system through a deep understanding ofthe mechanisms of reaction and deactivation; (2) to develop anefficient catalyst through the application of new zeoliticmaterials and the study of interplay between synthesis method,catalyst property, and reaction performance; (3) to scale up thesynthetic process of the catalyst with commercially available rawmaterials and establish standards to ensure the reliability and

repeatability for large scale catalyst production; (4) to find out asuitable reactor type with optimal reaction conditions andexplore the commercial availability; (5) to make a completeprocess flow diagram and set up a reliable reactor scale-upapproach; (6) to integrate the knowledge obtained from reactorscale-up with existing industrial experiences to make basicdesign package for commercial unit. There are many interestingresults and excellent reviews on MTO reaction and process,which, however, mainly focus on the fundamental re-search.2,3,6,20

The Dalian Institute of Chemical Physics (DICP) has beendedicated to the R&D of the MTO reaction for more than 30years, aiming at the development of a commercially availableprocess technology.21,22 Based on DICP’s MTO technology,namely, dimethyl ether or methanol-to-olefin (DMTO), theworld’s first MTO unit was constructed and started up inAugust 2010 in Baotou, China, which is considered as animportant milestone and critical step for producing light olefinsfrom coal. In this historical perspective, we would like to sharesome of our research and experiences in the development ofDMTO technology in DICP related to the aforementionedproblems, from the fundamental research, catalyst synthesis,and process development to commercialization. We wish thisperspective might shed some light on how the fundamentalresearch of catalysis in the laboratory, via the multidisciplinal

Received: January 4, 2015Revised: February 4, 2015Published: February 10, 2015

Perspective

pubs.acs.org/acscatalysis

© 2015 American Chemical Society 1922 DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

This is an open access article published under an ACS AuthorChoice License, which permitscopying and redistribution of the article or any adaptations for non-commercial purposes.

collaborations, can be eventually developed into a commercialtechnology.

2. HISTORICAL DEVELOPMENT OF MTO IN DICP

The MTO research in DICP began in 1982 as a projectsupported by Chinese government and Chinese Academy ofSciences due to the concerns about the shortage of oil supplybecause of the oil crisis in the late 1970s. Two research groupswere involved in the early researches, and one was in charge ofzeolite synthesis and the other focused on zeolite modificationand catalyst development. Based on ZSM-5 type catalyst, withparticipation of third research group, DICP finished a 300 t/aMTO fixed bed pilot plant test in 1993. In the meantime,following the findings of excellent MTO performance oversmall pore molecular sieves such as SAPOs and the idea thatsynthesis of dimethyl ether is more favorable in thermody-namics than that of methanol, we proposed a syngas viadimethyl-ether-to-olefins (SDTO) method in early 1990s.22 Inthis process, the modified SAPO-34 catalyst was used toconvert dimethyl ether to light olefins in a fluidized bed reactor.A pilot scale test on SDTO was finished in 1995. The SDTOresults indicated that they met the expectation; however, it wasstill less economically competitive compared to the oil route atthe time of low oil price. The MTO research in DICP thenfocused on molecular sieve synthesis and modifications in orderto further improve the catalyst performance, which led to thedevelopment of DMTO technology.In 2004, the DICP had a chance to demonstrate the MTO

process at a large scale by collaborating with SYN EnergyTechnology Corporation and SINOPEC Luoyang Petrochem-ical Engineering Corporation. The latter is a well-knownengineering company in China. DICP spent two years inscaling up the catalyst manufacturing and constructing ademonstration unit with a methanol feed rate of 16 kt/a.The successful running of the demonstration unit not only

provided the basic design data for the commercial units designbut also encouraged the rapid industrial applications of DMTOtechnology in China. Up to now, seven DMTO units have beencommissioned and put into stream.

3. FUNDAMENTAL STUDY ON MTO REACTION

MTO is an autocatalytic reaction,23,24 in which the initialformation of a small amount of products leads to an enhancedmethanol conversion until the efficient production period isreached. How C−C bond forms from C1 reactant has beendebated in the past decades. Early studies proposed many directmechanisms to explain C−C bond formation from methanol ordimethyl ether, such as carbene mechanism,1,25,26 oxysoniumylide mechanism,27,28 carboncation mechanism,24,29 free radicalmechanism,30,31 and so on, which were later proved energeti-cally unfavorable by theoretical calculations. Over fresh catalyst,the slow kinetic rate of the initial methanol conversion impliesthat C−C bond formation during this period possibly goesthrough the reaction route with relatively high energy barrier.Until now, how the first C−C bond generates is still acontroversial issue.Some work in the early 1980s proposed the indirect pathway

of olefins generation. Dessau and co-worker32,33 suggested thealkene-based pathway, and Mole and co-workers34,35 noted theprofound influence of aromatics and suggested the aromaticcocatalysis in the generation of olefins. In 1993−1996, thehydrocarbon pool (HCP) mechanism was proposed by Dahl

and Kolboe,34−38 i.e., cyclic organic species confined in thezeolite cage or intersection of channels act as cocatalysts for theassembly of olefins from methanol. The nature of HCP specieshas been investigated by several groups,3,4,20,39−42 and intensivestudies have proved that methylbenzenes are the HCP speciesin MTO reaction on SAPO-34 and revealed the correlationbetween the presence of methylbenzenes in SAPO-34 cagesand the formation of olefin products,5,43−48 in which Xu andWhite were the first to determine the structure of HCP speciesin an eight-membered ring molecular sieve catalyst.41,42 Withthe aid of solid-state NMR spectroscopy, some carbenium ionshave been identified as intermediates confined within zeolitecatalysts, such as 1,3-dimethylcyclopentadienyl cation, indanylcation, and 1,1,2,4,6-pentamethylbenzenium cation withinHZSM-5; heptamethylbenzenium cation within H-beta, andheptamethylcyclopentenyl cation within SAPO-34.49−53 How-ever, the formation of most of these carbenium ions duringmethanol conversion was proved via indirect ways. The directobservation of such HCP species and their roles in methanolconversion over zeolites or SAPO catalysts under real workingconditions is still a big challenge.Two distinct reaction routes, the side-chain methylation

mechanism and the paring mechanism, have been proposed toexplain the MTO reaction pathway according to the HCPmechanism.34,35,54,55 For the two reaction mechanisms,carbenium ions and their deprotonated counterparts areinvolved and act as the important reaction intermediates forolefin production.9,48,56−59 Specifically, the paring routeinvolves the contraction of six-membered ring ions (poly-methylbenzenium cations) and the expansion of five-memberedring ions (polymethylcyclopentenyl cations). In contrast, theside-chain methylation route proceeds via the methanolmethylation on polymethylbenzenium ions and subsequentelimination of side chain groups to produce olefins. Theobservation of carbenium ions and the explanation of the olefingeneration from such intermediates present two essentialquestions in HCP mechanism study.The aromatics-based hydrocarbon pool mechanism is

important for cage type SAPO and H-beta zeolite with requiredspace for the formation and function of the bulky HCPspecies.44,46−48,54,55 In 2006, dual-cycle concept was suggestedto explain the difference of the formation of ethene and higherolefins over H-ZSM-5.7,60,61 It is suggested that ethenegeneration over H-ZSM-5 is related to aromatic-based HCPmechanism, and the propene and higher olefins formationfollows an olefin methylation and cracking route.For the deactivation of MTO catalysts, external coke

formation is responsible for the deactivation of ZSM-5 bypoisoning active sites or blocking pores.62 Due to the narrowpore opening and big supercage, methanol conversion overSAPO-34 is characteristic of very quick coke formation andresidue in the cage. Haw and co-worker4 reported that duringmethanol conversion, methylbenzenes, the most activeconfined organics, formed in the cage of SAPO-34 areconverted with time on stream to methylnaphthalenes andpolyaromatics, phenanthrene derivatives and pyrene, the largestaromatic ring system that can be accommodated in thenanocages of the catalyst. The mass transport of reactantsand products will be greatly reduced with the accommodationof these bulky coke species, which cause the deactivation of thecatalyst.4 Considering the quick deactivation of SAPO-34, thefluidized-bed is applied in the industrial MTO process withreaction-regeneration cycle to keep the activity of catalyst.63

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1923

In the R&D history of DMTO technology, in parallel to theprocess development, the methanol conversion mechanism wasalso studied in order to enhance catalyst performance, optimizethe reaction conditions, and develop new highly selectivecatalysts.3.1. Methylation in MTO Reaction. Our study on

coreaction of ethene and methanol over different zeolites,SAPO-34, ZSM-22, ZSM-564−68 shows that there are at leastthree types of reactions involved in the coreaction system: thedirect conversion of methanol (MTO), the direct conversion ofethene, and the methylation of ethene by methanol. Thecompetitions among the three types of reactions are stronglyinfluenced by the catalyst acid properties and can also bealtered by catalyst modification. Complete suppression of directconversion of methanol or ethene can be achieved over ZSM-22 and modified ZSM-5.66,67 Using precoked ZSM-22 catalyst,on which the conversion of sole methanol or sole ethene is verylow, cofeeding methanol and ethene gives rise to enhancedmethanol conversion and ethene conversion, and at the sametime, propene selectivity attains to higher than 80% atprolonged reaction time (Table 1).66 Over hydrothermallytreated P, La modified ZSM-5 catalysts with extremely weakacidity, 12C-ethene and 13C-methanol are co-fed and theisotopic distribution of propene indicates that most propene(89%) contains one 13C atom from methanol, implying ethenemethylation by methanol is responsible for the enhancement ofconversion and propene generation in the coreaction system.64

The further methylation of propene generates butane-2 with 213C atoms from methanol.64 We also observe the same trendover SAPO-34, on which propene selectivity is also improved inthe co-feeding system.65,66 These results indicate that over themodified catalyst with weak acidity, the HCP mechanismcannot work, and therefore, alkene methylation and thecracking route will become the efficient and selective routefor olefin production. In the normal MTO process with highmethanol conversion, the methylation reaction should alwaysbe present, but it is implicit in the reaction system. Thisinteresting methylation reaction could be useful for higherolefin production, such as the production of propene from thecoreaction of ethene and methanol, and for adjusting theethene/propene ratio in the MTO reaction.

3.2. Observation of Important Carbenium Ions andTheir Role in Olefin Generation. Heptamethylbenzeniumcation (HeptaMB+) is of particular importance as a reactionintermediate in the MTO reaction,4,69 and its synthesis hasbeen made by the coreaction of benzene and methanol over H-beta, H-MCM-22, and H-mordenite.8,9,50 However, over themolecular sieves with narrow eight-membered ring poreopening such as SAPO-34, the coreaction of benzene andmethanol is unfeasible because of the diffusion limitations. Abreakthrough of HeptaMB+ observation was made by use of anewly synthesized catalyst DNL-6,70−72 a eight-membered ringSAPO molecular sieve with an RHO structure possessing largeα cages and relatively high acid concentration and strength,which can accommodate and stabilize the bulky carbeniumcations. For the first time, heptaMB+ is directly captured overDNL-6 by solid-state NMR during methanol conversion underreal working conditions (Figure 1), and this observation is

consolidated by GC−MS measurement of confined organics in

the catalyst.64,65 The 12C/13C-methanol switch experiments

confirm the higher reactivity of polymethylbenzene than that of

methylbenzenes with less methyl group substitution, implying

the important role of these intermediates in the olefin

formation.73,74

Table 1. Reaction Results of Methanol, Ethene, and Co-Feeding of Both over ZSM-2266

run no.a,b 1 2 3

feed/ml per minCH3OH 20 0 20C2H4 0 40 40He 40 20 0TOS/min 6 60 87 6 25 6 33 60outlet/C-mol %C2H4 12.3 0.23 0.22 98.7 98.8 61.8 73.8 77.5C3H6 29.0 0.10 0.06 0.11 0.09 14.1 5.62 2.35C4

+ 51.1 0.37 0.13 1.18 1.08 17.6 3.94 0.37CH3OH/CH3OCH3 3.47 98.5 99.2 - - 6.11 16.5 19.7C1

0-C30 4.13 0.80 0.42 0.04 0.03 0.31 0.16 0.11

CH3OH conv. %c 96.5 1.51 0.83 - - 71.0 21.7 6.60C2H4 conv. % - - - 1.33 1.21 21.6 6.51 1.78C3H6 sel. C-mol % 30.0 6.80 6.64 8.15 7.47 44.0 57.8 83.0

aT = 500 °C, ethene WHSV = 18 h−1, methanol WHSV = 10 h−1. bThree reaction runs were successively carried out without changing the catalystsample. In the first run, only methanol was fed. Then the catalyst bed was purged with helium and only ethene was fed (run 2). After the followinghelium flushing, the coreaction of ethene and methanol was conducted (run 3). cConversion was calculated with CH3OCH3 included.

Figure 1. 13C MAS and CP/MAS NMR spectra of the DNL-6 catalystafter 13C-methanol conversion at 275 °C for ∼50 min. * indicates thespinning sideband and Δ indicates the background.73

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1924

The difference of paring and side-chain methylationmechanism is whether the reaction route involves the ringcontraction of benzenium cations and the ring expansion ofcyclopentenyl cations or not. The isotopic switch experimentsoffer the evidence for the suggested reaction route (Figure 2).73

In the mass spectra of hexamethylbenzene (HMB) andhexamethylmethylenecyclohexadiene (HMMC, the deproto-nated form of heptaMB+), one notable observation is that thenumber of 13C atoms incorporated into HMB or HMMC is lessthan that of the substituted methyl groups (6 or 7). The mostincorporated 13C atoms are exactly the total count of methylgroups in the molecules. The other notable observation is thatthe M/Z gap between molecular ions of HMB (or HMMC)and the molecular ions with one methyl group lost is 16,indicating the location of all six 13C atoms of hexamethylben-zene (or seven 13C atoms of HMMC) in the methyl groupsrather than in the benzene ring. These two observations provethe feasibility of methanol conversion via side-chain methyl-ation mechanism, which involves a key step of methylation ofhexaMB to form heptaMB+ for further methylation andelimination to generate olefin products. (Figure 3).74

We also succeed in observing the simultaneous formation ofheptamethylbenzenium cation (heptaMB+) and pentamethyl-cyclopentenyl cation (pentaMCP+) under the real workingconditions of methanol conversion over the catalyst with CHAtopology,75 which made it possible to elucidate the reactionmechanism over the industrial DMTO catalyst. The observa-tions of heptaMB+ and pentaMCP+ and the correlationbetween their concentration and methanol conversion meanthat both of the two reaction routes, i.e., the paring route andside-chain methylation route, are possible for methanolconversion over the catalyst with CHA topology. Thetheoretical calculation predicted the reaction pathway withthe involvement of heptaMB+ and pentaMCP+ as theintermediates. Both of the two reaction routes for propeneproduction are energetically feasible, while the side-chainmethylation mechanism is predominant due to the lowerenergy barrier.

Using the techniques including in situ 13C NMR and GC-MS, we systematically investigated the confined intermediatesin SAPOs and other zeolites with the same eight-memberedring and different supercages.76,77 The results are summarizedin Table 2. These results indicate that the HCP mechanism isgenerally effective for the reaction; however, the carbeniumcations vary with the cavity size with some extent of diversity.In smaller cages, the cations are formed with less substitutedmethyl group or smaller molecular size. Although shape-selectivity is commonly believed to be effective for these eight-membered ring molecular sieves in the MTO reaction, cavitystill has great influence on selectivity considering the bigdifference of their product distributions. It seems that cavitycontrols the selectivity if the ring-opening of molecular sieveshas the same size.77

3.3. Deactivation at Low Reaction Temperature. Aninteresting phenomenon, as shown in Figure 4, is observed

Figure 2. Comparison of mass spectra of hexamethylbenzene (left) and HMMC (right) obtained in the isotopic switch experiments over DNL-6 (9μL of 13C-methanol injection after continuous flow 12C-methanol conversion for 60 min at 275 °C, WHSV of MeOH = 2.0 h−1, He/MeOH (in mol)= 3) with the mass spectra of 12C-HMB and HMMC (M) and simulated mass spectra of 13C- labeled HMB and HMMC (M′).73

Figure 3. Side-chain methylation mechanism for olefin generation overDNL-6.74

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1925

when we study the temperature-programmed methanolconversion in a microscale fluidized-bed reactor.78 When thereaction is fed at 250 °C, no methanol conversion occurs. Onlymethanol and dimethyl ether appear in the effluents until the

temperature gradually increases to 300 °C, showing an obviousinduction period of methanol reaction. At the temperaturerange of 300−325 °C, light olefins and other hydrocarbonsappear among the effluents, and the yield of hydrocarbonproducts increases with temperature. It is worthy to note theparticular and remarkable decline of methanol conversion inthe following period with temperature increase from 325 to 350°C. When the temperature is further increased to 400 °C, thecatalyst reactivity is partially recovered, and methanolconversion is improved again during this period. This particularobservation drives us to study the catalyst deactivation ofmethanol conversion at low temperature.Isothermal methanol conversion carried out at low reaction

temperature (300−350 °C) under the same condition presentsan obvious induction period and quick deactivation.79,80 Theinduction period is shortened with an increase in reactiontemperature. Diamondoid hydrocarbons largely form and areretained in the SAPO-34 catalyst as a kind of newly foundresidue hydrocarbon. Different from polymethylbenzenes, thereaction center of the MTO, 12C/13C methanol switchexperiment performed over the catalyst with confinedpolymethylbenzene and polymethyladamantanes confirms thatthe confined methyladamantane molecules are inactive whencontacting with the feed of methanol. The accumulation ofthese cyclic and saturated hydrocarbons in the catalyst blocksthe access of methanol to methylbenzenes and suppresses thesuccessive generation of the HCP species, accounting for thequick deactivation of SAPO-34 at 300 °C (Figure 5).79 On thebasis of this study carried out at low temperature underisothermal condition, we combine the confined coke speciesevolution with the catalytic performance of methanol reactionunder the condition of temperature-programmed increase. Inthe temperature range of 300−350 °C, beside the formation ofmethylbenzenes as the active reaction center, methyladaman-tanes are largely formed and cause the occurrence ofdeactivation at the temperature range of 325−350 °C. Withthe temperature increase, methylbenzenes and methyladaman-tanes are transformed to methylnaphthalenes and polycyclicaromatics at higher temperature. The evolution of the confinedorganics corresponds to the initial reactivity enhancement ofthe catalyst at temperature range of 300−325 °C, the loweredreactivity at 325−350 °C, and recovered methanol conversionat 350−400 °C (Figure 4).78 Some extended work indicates

Table 2. Carbenium Ions Observation in Cage-Type Eight-Membered Ring Zeolite Catalysts

Figure 4. Effluent distribution of fluidized-bed methanol conversionunder the condition of programmed increase of temperature.

Figure 5. GC-MS analyses (left) of confined organics after methanol conversion at 300 °C for 17 (a), 32 (b), 47 (c), 62 (d), and 92 min (e);Methanol conversion (middle) and methylbenzenes and methyladamantanes (right) variation with time on stream. * internal standard.79

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1926

adamantane hydrocarbons can also be formed and accom-modated in other eight-membered ring and cage-type SAPOmolecular sieves with AEI and LEV topology. These cage typemolecular sieves provide the catalytic surrounding for theformation of adamantane hydrocarbons as confined products inmethanol conversion under the reaction condition of lowtemperature.During methanol conversion, coke deposition over the

catalyst has been also studied with TEOM (tapered elementoscillation microbalance) under the real reaction condition.80,81

The coking rate is quite different at various reactiontemperature. High- (500 °C) and low-temperature (300 °C)reactions present higher coking rate than the reactionsperformed under mild reaction condition (400−450 °C).80

The extremely high coking rate at low reaction temperaturecorresponds to the quick deactivation during methanolconversion with predominant adamantane hydrocarbonsformation and accommodation in the catalyst.3.4. Reaction Network. Our studies of methanol

conversion over zeolites and SAPO catalysts with differenttopology indicate that the aromatics-based HCP reactionmechanism is dominant for the MTO reaction over SAPO-34than on ZSM-5 and ZSM-22.65−68 Dual-cycle reactions proceedover ZSM-5 catalyst, and olefins methylation and crackingmechanism occurs in the reaction over ZSM-22.66−68 Even thesignificant contribution of the HCP mechanism for cage-typeSAPO catalysts, olefin methylation and cracking mechanismalso works in some cases. The acid density variation playsdifferent roles in the alkenes-based cycle and aromatics-basedcycle. When methanol conversion to light olefins takes placeover the eight-membered ring AlPO-18 and SAPO-18 with AEIcavity, the catalyst with high-density Brønsted acid sites canaccelerate the formation and accumulation of cyclic organicspecies over the catalysts. The aromatics-based HCPmechanism plays an important role for methanol conversionover SAPO-18 with more Brønsted acid sites; however, overAlPO-18, because the carbenium ions cannot be formed overnonacidic catalyst, the conversion is very low and, theformation of olefins mainly follows olefin methylation andcracking route.82

In the actual methanol reaction over acid zeolites or SAPOcatalysts, considering the occurrence of all the possiblecompetitive reactions, including the olefin formation reactionthrough aromatics-based or olefins-based reaction route (i.e.,olefin methylation, olefin oligomerization, and cyclization,olefin cracking, and other reactions which may causedeactivation), a reaction network is proposed and presents acomplicated reaction system (Figure 6), in which somereactants and products are reciprocal or overlapped in a seriesof reactions following different reaction routes. There are twomain reaction modes in the network related to coke formation,one by further transformation of HCP intermediates and theother one by further conversion of olefin products without

involving HCP intermediates. In principle, the first cokingreaction can be restricted by selecting molecular sieve catalystwith suitable cavity size, which could be a possible way toenhance the catalyst lifetime.

4. SYNTHESIS AND CRYSTALLIZATION MECHANISMOF SAPO MOLECULAR SIEVES

Silicoaluminophosphate (SAPO) molecular sieves were firstinvented by the scientists at Union Carbide Corporation in1984.83,84 Among the SAPOs, SAPO-34 with the CHAstructure is particularly interesting and demonstrates excellentcatalytic performance in the MTO reaction due to thecontribution of small pore, medium acidity, and highthermal/hydrothermal stability.15,21,85,86 The CHA frameworktopology, characterized by cylinder-like cages with eight-membered ring openings, has been evidenced to be the idealbreeding ground for hydrocarbon pool intermediates in theMTO reaction.36,48

Hitherto, many research works have been reported on thesynthesis, physicochemical and catalytic properties of SAPO-34.87−100 DICP has also dedicated great efforts to the synthesisand crystallization mechanism of SAPO-34, as well as therelationship among the Si coordination environment (acidity),Si distribution, and the MTO catalytic performance. In the late1980s, DICP first reported the excellent catalytic performanceof SAPO-34 in the MTO reaction and its good regenerationstability.85,101 This exciting finding was further strengthened byour following investigations on the hydrothermal stability ofSAPO-34, which kept more than 80% relative crystallinity after100% steam treatment at 800 °C for 45 h.102 These resultsconfirmed that SAPO-34 could be taken as a promising MTOcatalyst for industrial application. In the subsequent researches,much attention was paid to explore new templates for theSAPO-34 synthesis, considering that templates may influencethe morphologies, microstructures, and acidic properties, andthus lead to optimized catalytic properties. We found that, as anexample, triethylamine (TEA), diethylamine (DEA), and theirmixtures could direct the crystallization of SAPO-34 with goodcatalytic results.103−105 The observations were attractive fromthe practical perspective because these amines were muchcheaper than the traditional tetraethylammonium hydroxide(TEAOH), and their use would greatly reduce the cost formanufacture of the catalyst. Subsequently, a symmetricinvestigation on the relationship of syntheses, Si coordinationenvironments, and catalytic properties were carried out.106−109

It was recognized that SAPO-34 with lower Si content,corresponding to the major existence of Si(4Al) species inthe framework (medium/strong acidity), generally had higherolefins’ selectivity and longer lifetime. As shown in Figure 7,with the rising of Si content in the SAPO-34, the Sicoordination environments become complex, and the corre-sponding acid strength/concentration increases. SAPO-34(a)catalyst with low Si content and single Si(4Al) species exhibitlong lifetime and low hydrogen transfer index (correspondingto the low selectivity of propane and slow coking rate). Furtherstudies then focused on the crystallization process of SAPOmolecular sieves to understand the mechanisms governing thenucleation, crystallization, and Si distribution for a bettercontrol of composition and Si environmental of the syntheticproduct.110−113 The insight on the crystallization mechanismshas been proved to be an effective way to optimize the synthesisof SAPO-34 with improved catalytic properties.Figure 6. Reaction network of methanol conversion.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1927

In the early research stage, when the understanding ofmolecular sieves was relatively poor, it was difficult todetermine which molecular sieve could best fit the specifiedapplication because many small-pore SAPO molecular sieves,such as SAPO-44, -17, -18, -35, -56, and so forth, could bepotentially used as shape-selective catalysts as far as the poresize openings are concerned. We synthesized all SAPOsmentioned above and explored their MTO performances.114,115

Unfortunately, all of them exhibited inferior catalytic perform-ance (both the lifetime and selectivity to ethene plus propene)to SAPO-34. Furthermore, various MeAPSOs and metal-modified SAPO molecular sieves including SAPO-34 werealso investigated, and some of them showed obvious improve-ment in the selectivity to olefins.114,116,117 However, in mostcases, the increasing amount of CO and CO2 could be found inthe reaction product due to the existence of transitional metalsthat may cause the decomposition of methanol.

4.1. CRYSTALLIZATION MECHANISM AND SIDISTRIBUTION IN THE CRYSTALS

Crystallization mechanism is an interesting issue in the study ofmolecular sieves, which can offer a more rational approach tothe design and synthesis of molecular sieves with optimizedacidic/catalytic properties.Crystallization of SAPO-34 was first studied with TEA

template at 200 °C from a gel composition of 3TEA:0.6-SiO2:Al2O3:0.8P2O5:50H2O.

110 Figure 8 illustrates the pro-posed model of SAPO-34 crystallization and the Siincorporation mechanism. It is found that preliminary structureunits similar to the framework of SAPO-34 have already formedbefore the crystallization start (0 h and low temperature), andthe nucleation of SAPO-34 comes from the structurerearrangement of the initial gel and the condensation of the

hydroxyls. The solution support is important for furthercrystallization of SAPO-34. Regarding the Si incorporation, itreveals that Si atoms directly take part in the formation of thecrystals. At the early stage of the crystallization (<2.5 h), a SM2mechanism occurs (substitution of P by Si), generating aSi(4Al) environment. Subsequently, the relative content of Si inthe products increased slowly, and Si environments becomecomplex (Si(nAl), n = 0−4), suggesting the co-occurrence ofSM2 and SM3 (substitution of P−Al pair by 2Si) mechanisms.Further investigation on the crystallization of SAPO-34

templated by DEA (2DEA:0.6SiO2:Al2O3:0.8P2O5:50H2O, 200°C) gives similar conclusion as that by TEA system about thecrystal formation/growth and Si incorporation mechanism.118

Higher Si incorporation amount into the framework wasobserved with DEA template, possibly due to the smallermolecule size of DEA. This is also consistent with the higherDEA molecular density in the CHA cage (1.75 DEA vs 1.0 TEAper cage). Notably, a nonuniform distribution of Si atoms in thecrystals was revealed, which shows a gradual increase from thecore to the surface (Figure 9). Given that the acid strengthcorresponding to different Si species in SAPO molecular sieveshas the order of Si(1Al) > Si(2Al) > Si(3Al) > Si(4Al),119 this

Figure 7. NH3-TPD profiles (top-left) and 29Si MAS NMR spectra (top-right) of SAPO-34s and their MTO catalytic results (bottom-left). SAPO-34was synthesized with TEA template (3TEA:1Al2O3:1P2O5:xSiO2:50H2O, 200 °C, 48 h). a: x = 0.2; b: x = 0.4; c: x = 0.6. Reaction conditions: T =450 °C, WHSV = 1.2 h−1, 30 wt % methanol solution. Hydrogen transfer index (HTI) over the catalysts is shown in the bottom-right of the figure.

Figure 8. Model of the SAPO-34 crystallization process and the Siincorporation mechanism.110

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1928

finding suggests that the acidic sites along the crystals are notuniformly distributed and the stronger acidic sites are morelikely concentrated on the external surface. As shape-selectivereaction always causes diffusion effect, it is reasonable tospeculate that the stronger acidic sites on the external surfacewould have high priority for the reaction and cause lessselectivity and more coke formation.120 The larger cokecompounds formed in the cages at the edges of the crystalsmay further cause diffusion limitation and inhibit the release ofproduct generated inside. It implies that the surface Sienrichment phenomenon is detrimental to the MTO reaction,which should be avoided for a better shape selectivity of theMTO reaction.Besides SAPO-34, three other SAPO molecular sieves

(SAPO-35, -11, and -5)111,112 with different topologicalstructures were also investigated with respect to theirhydrothermal crystallization process, in order to gain acomprehensive understanding of the synthesis mechanisms.Although the concrete crystallization process of each SAPOmolecular sieve is different from others, there are still somecommon features that could be concluded as follows: (1) thecrystal growth follows a solution-mediated transport mecha-nism; (2) Si atoms incorporate into the SAPO frameworksfrom very beginning of the crystallization and their content inthe products increase gradually; (3) the silica enrichment onthe crystal surface occurs for all SAPO samples investigated(Table 3), and the extent is influenced by the framework typeand synthetic conditions.

4.2. Synthesis of SAPO-34 with Single Si(4Al) Environ-ment. One simple way to control the Si coordinationenvironment in SAPO-34 is to vary the Si content in thestarting gel. However, the synthesis of SAPO-34 with only theSi(4Al) environment is difficult because crystal impurities mayform together with SAPO-34 when employing low Si content.Crystallization conditions and starting materials should be

carefully controlled to achieve a successful synthesis.118,121,122

According to our crystallization research, the initial gel withmedium/high Si content and shortened crystallization timewould also provide an easy approach to produce SAPO-34 withsingle Si(4Al) species, which, however has a low product yield.Fluorine-mediated synthesis has been widely employed in

zeolite crystallization due to the effective mineralization abilityof F− ions. It can form chelates with the reaction sources (Al/P/Si) and promote the dissolution of pseudoboehmite andsilica particles. With the assistance of fluoride, crystals withfewer framework defects could be obtained. Table 4 displays

the product compositions, Si(4Al)/Si(3Al) ratios, and MTOresults of SAPO-34-F (HF/Al2O3 = 0.33 in the initial gel) andconventional SAPO-34.123 Dominant Si(4Al) species appear inthe SAPO-34-F sample, giving longer lifetime and lower alkaneselectivity, although the two samples show similar productcompositions. Given that SAPO-34-F possesses similar texturalproperties to conventional SAPO-34, it is reasonable toattribute the improved catalytic performance of SAPO-34-Fto its optimized Si coordination environments and further tothe corresponding acidity. Moreover, it is speculated that theformation of fluoride-Si chelates would lead to a relativelyhomogeneous release of Si during the crystallization and thusreduce Si enrichment on the external surface of SAPO-34crystals, as compared to the case without fluoride.In an attempt to lighten the Si enrichment on the external

surface, a facile postsynthesis modification method wasdeveloped by treating as-synthesized SAPO-34 with HF orNH4F solution.107 As shown in Figure 10, both the relativecrystallinity and Si content of the sample decrease following theincreasing HF concentration. In the meantime, the externalsurface of SAPO-34 becomes rather rough, and Si species otherthan Si(4Al) gradually diminish. This implies that F− ionsselectively remove Si atoms in the Si-rich regions of the crystalsurface. All modified samples exhibit prolonged lifetime with amaximum over SAPO-34−0.1HF sample (0.1 mol/L HFsolution used for the post-treatment).

4.3. Novel Methods To Synthesize SAPO-34. Asolvothermal synthesis route designated as aminothermalsynthesis, in which organic amines are used as both the

Figure 9. Schematic diagram of the crystallization process of SAPO-34with Si distribution in the crystals. The dot density in the square standsfor the Si concentration in the crystals.

Table 3. Bulk and Surface Compositions of SAPO MolecularSieves

elemental composition (mol %)

samplea XRF XPS Sisurface/Sibulk

SAPO-34-T Si0.080Al0.486P0.434 Si0.206Al0.453P0.341 2.58SAPO-34-D Si0.164Al0.478P0.358 Si0.231Al0.440P0.329 1.41SAPO-5 Si0.073Al0.488P0.439 Si0.155Al0.420P0.424 2.12SAPO-11 Si0.045Al0.498P0.458 Si0.224Al0.462P0.314 4.98SAPO-35 Si0.122Al0.488P0.390 Si0.187Al0.478P0.335 1.53

aData from ref 111. SAPO-34-T and SAPO-34-D were synthesizedwith TEA and DEA as the template, respectively.

Table 4. MTO Reaction Results over SAPO-34 Synthesizedwith and without HFa

sample SAPO-34 SAPO-34-F

product composition Si0.10Al0.49P0.41 Si0.09Al0.50P0.41Si(4Al)/Si(3Al)b 2.75 66.35lifetime (min)c 120−140 180−200selectivity (wt %) CH4 0.93 0.89

C2H4 31.29 35.66C2H6 0.84 0.69C3H6 39.55 38.41C3H8 7.04 5.03C4 11.66 10.83C5

+ 8.69 8.49C2

=+C3= 70.84 74.07

aReaction conditions: T = 450 °C, WHSV = 2h−1, N2 as carrier gas,TOS = 20 min. Gel compositions for the synthesis are3TEA:1Al2O3:1P2O5:0.6SiO2:xHF:50H2O (200 °C, 12h). x = 0 and0.33 for SAPO-34 and SAPO-34-F respectively; bDetermined from29Si MAS NMR spectra; cReaction duration with 100% methanolconversion.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1929

dispersing medium and the template, has been developed forthe synthesis of SAPO-34.124−126 The main features associatedwith this method are the high solid yield and fast crystallizationrate, which is important from the practical perspective. Forexample, 90% and 96.2% yields are acquired after 12 h at 200°C on the basis of N,N,N′,N′-tetramethylethylenediamine(TMEDA) and di-iso-propylamine (DIPA) systems, respec-tively. The less solubility of inorganic resources in theaminothermal system should be the main reason for theincreased solid product yield. Investigation on the amino-thermal crystallization process revealed that at first a lamellarAlPO material is formed, which promotes the dissolution ofresources and leads to a faster crystallization rate. Significantly,SAPO-34 templated by DIPA shows good MTO catalyticperformance with an ethene plus propene selectivity of 85.8%.Further extending this method to other amine systems, severalnovel templates for the synthesis of SAPO-34 have beendiscovered, such as diisopropanolamine, diglycolamine, and N-methyldiethanolamine. Moreover, a phase-transformationmethod has been proposed to synthesize SAPO-34 by usingother SAPO molecular sieves as precursors.72 This approachprovides a possibility to transform crystals containing SAPOimpurities to pure SAPO-34 product.The synthesis of nanosized or hierarchical SAPO-34 has

attracted great attention because it can reduce the diffusivelimitation and lead to an improved MTO catalytic lifetime.Many strategies have been reported, and most of them involvethe use of expensive TEAOH as the template.13,127−132 A top-down route (postsynthesis, milling, and recrystallization) hasbeen developed to prepare SAPO-34 nanocrystals by usingTEA template.133 The obtained nano-SAPO-34 showsconsiderably improved catalytic performance (Figure 11),resulting from a combination of shortened diffusion paths,decreased acid concentration, and reduced Si enrichment onthe crystal surface. This method is further proven to begenerally applicable to prepare other SAPO nanocrystals.Furthermore, hierarchical SAPO-34 has also been synthesizedusing organosilane surfactant as part of silica source andmesoporogen with TEA or DEA as a micropore template.134

5. CATALYST AND SCALE-UP

5.1. Scale-Up Synthesis of SAPO-34. DICP first scaledup the synthesis process of SAPO-34 in 1993 by employing a 1M3 autoclave, aimed to supply catalyst (named as DO123) forthe pilot-scale SDTO test. In 2005, a pilot catalyst synthesisplant with two 2 M3 autoclaves, which was close to acommercial scale to some extent, was built up. The purposesof the scale-up of the syntheses process were as follows: (1) to

supply molecular sieve for the production of DMTO catalystused for the demonstration unit; (2) to verify thereproducibility and reliability of the synthesis technology; (3)to find and solve problems that may exist in the large scalecatalyst manufacture such as heating rate, mixing homogeneity(stirring), and solid−liquid separation; (4) to further optimizethe synthetic conditions and reduce cost; (5) to establish anentire catalyst quality control system for the synthesis process.On the basis of the data and experiences from the pilotsynthesis plant, the industrial production of SAPO-34 in a largescale was realized in 2008.

5.1.1. Synthesis Planning. The choice of a synthesisprocedure suitable for industrial catalyst manufacture shouldconsider many factors. Good catalytic performance in MTOreaction is of particular importance. Meanwhile, solid yield,crystallization rate, and operational feasibility should also betaken into account. The synthesis of SAPO-34 developed byDICP employs TEA, DEA as the templates. The product hasthe characteristics of rapid crystallization, adjustable acidity, andparticle size. More importantly, the present template has alower boiling point (below 100 °C) than water. After thesynthesis, the amine could be easily separated from the motherliquid through flash distillation and reused for the synthesis,which could reduce the cost for synthesis and the followingwaste treatment. It is worth noting that all raw materials used in

Figure 10. 29Si MAS NMR spectra, relative crystallinity, product compositions, and MTO reaction lifetime of SAPO-34 before and after treatmentwith HF solution (RT, 12h). Reaction conditions: T = 450 °C, WHSV = 2h−1, N2 as carrier gas.

Figure 11. SEM images and MTO results of SAPO-34 before and afterpostsynthesis milling and recrystallization. Reaction conditions: T =450 °C, WHSV = 4 h−1, 40% methanol solution.133

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1930

the synthesis scale-up are commercially available, and onlythose who could provide a long-term supply with steady qualitycan be selected as the suppliers.5.1.2. Process of the SAPO-34 Scale-Up Synthesis. The

process flow diagram of the synthesis is shown in Figure 12.

The synthetic procedure is as follows: the Si/Al/P/templateresources were mixed with deionized water according to apredetermined sequence and stirred to achieve a homogeneousgel. The autoclave is then sealed and heated to thecrystallization temperature. After the crystallization is complete,the autoclave is quenched by cooling medium to 90 °C, atemperature at which the residue of organic amine was distilledout and collected. When the autoclave is further cooled tobelow 50 °C, solid products can be recovered by filtration,washing with deionized water, and drying in the air.The recovery of amine during the cooling period is important

from the economic and ecological viewpoints, which couldimprove the environment of subsequent filtration unit andreduce the liquid waste disposal. After GC-MS analysis of thecollected amine solution, it can be readily recycled for thesynthesis. The obtained SAPO-34 shows good crystallinity andcatalytic performance, which is almost identical to the productsynthesized with fresh amine. Moreover, a large amount ofwaste liquid generated from the synthesis (filtration/washing),containing unreacted chemicals, can also be reused afteranalyzing the solid content. The synthesis with recycled liquidhas rapid crystallization rate and higher solid yield. This is likelydue to the small amount of crystals/nuclei (seeds) in therecycled liquid. These approaches reduce the emission of wasteand the cost of synthetic process, and make the process moreenvironmental friendly.5.1.3. Quality Control of SAPO-34 Synthesis. Establishment

of the quality control system and standard test method/procedure for the synthesis process is also important to keepthe product quality and maintain a steady production. Figure 12illustrates the quality control points of the SAPO-34 synthesisprocess (A1−A5). A1 is about the analyses of incoming rawmaterials to avoid the use of unqualified materials. A2 measuresthe crystal phase and purity of the solid in the synthesis slurry.A3 and A4 analyze the content of the distilled amine solution

and collected waste liquid, respectively, for their recycling use.A5 is the analyses of the SAPO-34 product including crystalphase, composition, morphology, textural properties, solidcontent, and catalytic performance. The qualified productwould be delivered to the next unit for the microsphere catalystpreparation.

5.2. Catalyst Manufacturing Process Scale-Up. TheDMTO catalyst with SAPO-34 as the active component isprepared by spray drying method and has the followingcharacteristics: microsphere, particle size distribution similar toFCC catalyst, high mechanical strength (low attrition index),and good catalytic performance.135−137 The manufacturingscale-up of DMTO catalyst was carried out in 2005, and ca. 25tons of catalyst was successfully produced and supplied for theDMTO demonstration unit (16 kt/a MeOH feed). Sub-sequently, following the licensing of DMTO catalyst technol-ogy, a 2000 t/a DMTO catalyst plant (containing molecularsieve synthesis and catalyst manufacturing) was constructedand put into production in 2008.

5.2.1. Process Flow of the Catalyst Manufacturing. Theprocess flow diagram of the catalyst manufacturing is illustratedin Figure 12. The manufacturing procedure is as follows:SAPO-34, binder, auxiliary agent, porogen, and deionized waterare mixed under stirring and milled to achieve sufficientuniform slurry. The waste liquid generated in the synthesissection could substitute part of water for the catalystmanufacturing. The well-milled slurry is then fed to spraydrier, by which microsphere catalyst is produced. Following asuccessive calcination, packaging, and storage, the final catalystcan be obtained.

5.2.2. Quality Control for Catalyst Manufacturing. Thequality control points of the catalyst manufacturing processhave been labeled in Figure 12 as CA1−CA4. CA1 is theanalysis of the raw materials except the SAPO-34 molecularsieves. CA2 detects the particle size distribution of the milledslurry to ensure the effective dispersion/mixing of thecomponents. CA3 measures the particle size distribution andmorphology of the semiproduct, which provides feedback tothe spray drying to keep the suitable particle size distributionfor product. CA4 is the analysis of the calcined product,including the solid content, composition, morphology, particlesize distribution, attrition index, textural properties, bulkdensity, and catalytic performance. Typical properties of theDMTO catalyst have been listed in Table 5. The typicalmorphology and particle size distribution is shown in Figure 13.

6. DMTO PROCESS DEVELOPMENTS6.1. DMTO Reactor Selection. The MTO reaction over

SAPO-34 catalyst is highly exothermal, and the heat of reactionis about −196 kcal/kg methanol feed when the reactiontemperature is at 495 °C. The reaction heat must be removedfrom the reaction bed simultaneously to keep the reactiontemperature in the designed range. In addition to that, theMTO reaction over SAPO-34 catalyst is featured by rapidcatalyst deactivation due to the coke deposition. The methanolconversion shows a rapid drop after the coke amount on thecatalyst reaching up to a maximum value (Figure 14). In orderto maintain high activity of catalyst in the reactor, the in-linecombustion of coke is required. At early stage, DICP hadinvestigated the fixed bed reactor for MTO process.Subsequently, the fluidized bed reactor was considered superiorto fixed bed reactor because of the excellent heat transferperformance and good fluidity of catalyst in fluidization state.

Figure 12. Process flow diagrams of the SAPO-34 synthesis (top) andcatalyst preparation (bottom). The quality control points of eachprocess have been labeled in the diagrams as A1−A5 and CA1−CA4.WD refers to waste disposal.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1931

On one hand, the great quantity of small catalyst particles canprovide a very large surface area, and the heat transfer is rarely aproblem in the fluidized bed. The temperature difference in thefluidized bed can be readily reduced to just several degrees viaoptimal operation. The heat removal is actually a commonpractice in fluidized bed processes in industries. One example isthe catalyst coolers widely used in fluid catalytic cracking(FCC) units in modern refineries. On the other hand, thecatalyst, when fluidized, can be transported between the reactor

and regenerator like a liquid, which makes the continuousregeneration of catalyst in DMTO process possible. In theDMTO process, the fluidized bed reactor−regeneratorconfiguration was adopted.ExxonMobil and UOP have applied for many patents on the

riser18 and fast fluidized bed reactor138 for the MTO process. Inour early studies, different types of fluidized bed reactors wereproposed and tested for DMTO process, which include twodowner reactors and a dense phase fluidized bed reactor. It wasfound that the catalyst deposited with a certain amount of cokecould favor the selectivity to ethene and propene even at highmethanol conversion (>99%), as illustrated in Figure 14. Anoptimal operation window exists near the coke content of ∼8%.To achieve an average coke content of ∼8% in the reactor,catalyst residence time should be around 60 min subject to thereaction temperature and WHSV. In this connection, both theriser and downer reactors seem not suitable for DMTO processas the catalyst residence time is only several seconds, and theoptimal operation window is hard to achieve without recyclingthe catalyst to accumulate coke on the catalyst. Furthermore,the methanol conversion is relatively low in either riser ordowner reactor. In the DMTO process, the dense phasefluidized bed reactor is chosen.Experiments in the laboratory showed that a catalyst−gas

contact time of about 2−3 s is necessary in order to avoidundesired byproduct in the DMTO reaction. Such a shortcatalyst−gas contact time means that an industrial DMTOreactor has to be operated at a gas velocity higher than 1 m/s,which otherwise leads to a fluidized bed with bed height lessthan 3 m. This imposes a great challenge for engineering designas it is extremely difficult, if not impossible, to arrange gasdistributor, cyclone diplegs, catalyst draw-off bin, and otherinternals in a space lower than 3 m. Actually, a high gas velocitycan result in an increased gas throughput. DMTO catalystparticles have similar physical properties as FCC catalyst, as willbe addressed in the next section, and are typically type Aparticles according to Geldart’s classification.139 For this type ofparticle, a superficial gas velocity of 0.5−1.5 m/s would makethe fluidized bed operating in the turbulent fluidized bedregime. An even higher superficial gas velocity will make thefluidized bed transit from turbulent fluidization to fastfluidization. A turbulent fluidized bed reactor offers apparentoperational advantages over the bubbling and fast fluidized beddue to the excellent catalyst−gas contact, high mass transferefficiency, and large solids holdup.140 In the DMTO process, aturbulent fluidized bed reactor operating at a superficial gasvelocity of 1−1.5 m/s is designed.

6.2. DMTO Fluidized Bed Reactor Scale-Up. The scale-up of a new fluidized bed process is rather an engineeringpractice than an exact science.141−143 First, the fluidizationbehavior in the reactor can vary significantly when the physicalproperties (i.e., size and density) of catalyst particles aredifferent.139 Second, the hydrodynamics in the fluidized bedsuch as the gas bubbles, solid mixing and dispersion, and gasmixing and dispersion is closely related to the bed diameter aswell as the operation conditions.144,145 The dream of chemicalengineers is to directly scale up the fluidized bed process by useof a reliable mathematical model.141 Although great effort hasbeen devoted to establish such a model141 in recent years, theDMTO fluidized bed reactor was scaled up through theexperiments at different scales.

6.2.1. Comparison between DMTO and FCC Process. Thefluidized bed reactor−regenerator configuration in DMTO

Table 5. Typical Properties of the DMTO Catalyst

BET surface area (m2/g) ≥180pore volume (cm3/g) ≥0.15attrition index (%/h) ≤2density (g/cm3) bulk density 0.6∼0.8

dense packing density 0.7∼0.9particle density 1.5∼1.8

particle size distribution (%)a ≤20 μm ≤520∼40 μm ≤1040∼80 μm 30∼5080∼110 μm 10∼30110∼150 μm 10∼30≥150 μm ≤20

catalytic performanceb lifetime (min) ≥120optimal selectivity of C2H4 plusC3H6 (wt %)

≥86.5

aMeasured with laser particle sizer. bTest conditions: fixed-fluid bed,10 g of catalyst, 450 °C, WHSV = 1.5 h−1, 40 wt % methanol solution.The lifetime refers to the reaction duration of conversion >98%(methanol and dimethyl ether considered as reactant).

Figure 13. SEM image and particle size distribution of DMTOcatalyst.

Figure 14. Effect of coke deposition on the methanol conversion andselectivity to light olefins in microscale DMTO fluidized bed reactor at450 °C.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1932

process is very similar to that in the FCC process. We hadmade a detailed comparison between the DMTO and FCCprocess in terms of feedstock, reaction mechanism, catalyst,product, and process (see Table S1). The big difference is thatin the FCC process, the reactor is a riser, while in DMTO, it isa turbulent fluidized bed. However, in the FCC process, theregenerator is operating in the turbulent fluidized bed regime.The experience in the design and operation of FCC turbulentfluidized bed regenerator can certainly help to reduce the risk ofDMTO reactor scale-up. The good fluidization performance ofFCC catalyst is well-known to chemical engineers for decades.The FCC catalyst is a type A particle according to Geldart’sclassification.139 Therefore, it had been decided that theDMTO catalyst would be a type A particle and has similarphysical properties as the FCC catalyst. Typical properties ofDMTO catalyst are shown in Table 5. The two most importantparameters, the particle size distribution and particle density,were specifically designed during the catalyst manufacturing.The percentage of fines less than 40 μm in the catalyst, which iscritical for catalyst circulation, is maintained at 15%.6.2.2. DMTO Reactor Scale-Up. To be sufficiently confident

in designing a commercial DMTO reactor, the followingproblems have to be resolved: (1) fluidization of catalyst inmethanol gas; (2) catalyst circulation; (3) catalyst stripping;(4) catalyst attrition; (5) optimal catalyst residence time; (6)optimal gas−catalyst contact time; (7) effect of reactor size onfluidization; (8) effect of reactor size on reaction; (9) heat ofreaction.Figure 15 shows the process of DMTO fluidized bed reactor

scale-up. Table 6 summaries the typical operation conditions at

different scale. The microscale fluidized bed is a fixed fluidizedbed operated at bubbling fluidization regime without catalystcirculation. In this stage, the main task is to evaluateperformance of DMTO catalyst and investigate the optimaloperation window. It was found that the coke content is nearlya linear function of catalyst residence time (RT). Wedetermined the optimal catalyst RT in microscale fluidizedbed reactor, which was consequently implemented into thedesign of pilot-scale fluidized bed reactor. Also the microscale

reactor was used to study the reaction kinetics of DMTOprocess. Pilot-scale reactor is connected to a fluidized bedregenerator via standpipes, thus a complete catalyst circulationbetween reactor and regenerator can be tested. The pilot-scalereactor was operated at bubbling fluidization regime, and thecatalyst circulation rate was determined by the catalyst RTobtained from the microscale experiments. In the pilot-scalereactor, the catalyst stripping was also tested. The gas−catalystcontact time and WHSV were kept the same for bothmicroscale and pilot-scale reactors, which are 2 to 3 s and 2−3 h−1, respectively. The fluidization performance of DMTOcatalyst was preliminarily demonstrated in the pilot-scalereactor. The methanol conversion is above 99% for bothmicroscale and pilot-scale reactors. Typical selectivity (innonaqueous products) to ethene and propene is 80% in bothmicroscale reactor and pilot-scale reactor.In the microscale and pilot-scale, the superficial gas velocity

in the reactor is 1 to 20 cm/s, which lies in the bubblingfluidization regime. In the demonstration scale, the superficialgas velocity in the reactor was promoted to achieve a trulyturbulent fluidization. In the demonstration unit, of particularimportance is the high WHSV (∼5 h−1) and superficial gasvelocity (1∼1.5 m/s), which was significantly different fromthat in microscale and pilot-scale reactor. In addition to that,the catalyst circulation, stripping, attrition, as well as heat ofreaction were carefully studied in the demonstration unit.

6.3. DMTO Demonstration Unit. In 2006, the DMTOdemonstration scale unit was constructed in Huaxian, Shanxi.Figure 16 is the schematic flow diagram of the DMTO

demonstration unit. The unit consists of fluidized bed reactor,fluidized bed regenerator, quench tower, and sour waterstripper. The aims of the demonstration unit are to study thehydrodynamics and reaction performance of DMTO turbulentfluidized bed reactor and provide data for commercial reactordesign, in addition to get exact product distribution data

Figure 15. Scale up of DMTO fluidized bed reactor.

Table 6. Scale-Up of DMTO Fluidized Bed Reactor

microscale pilot-scale demo-scale commercial-scale

MeOH feed, kg/d 0.5 50 50 000 5 500 000catalyst inventory, kg 0.01 1 400 45 000superficial gas velocity, m/s 0.01∼0.1 0.01∼0.2 1∼2 1∼2fluidization regime bubbling bubbling turbulent turbulent

Figure 16. Schematic flow diagram of the DMTO demonstration unit.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1933

including all trace amounts of impurities, such as oxygenates,alkynes and dienes, for the following olefin recovery unitdesign. The DMTO reactor in the demonstration unit has adiameter of 1 m, with a methanol feed rate of 2.5 t/h. The unitwas running for a total of 1200 h. The influences of theoperation conditions on methanol conversion and selectivity toethene and propene were studied. Catalyst attrition was alsotested in the demonstration unit. Figure 17 shows the results

during continuous operation, and Table 7 summarizes theaverage reaction results. In the demonstration unit, the averageselectivity to ethene and propene was 79.21%, and themethanol conversion was higher than 99%. Furthermore, theethene/propene ratio could be adjusted in the range of 0.8 to1.2 by varying operation conditions. The data from thisdemonstration unit was implemented in the design package ofcommercial DMTO units.

6.4. Reaction Engineering Research for DMTOProcess. DICP has carried out a series of reaction engineeringresearch for DMTO process, which aims at understanding thefluidization of the DMTO catalyst, optimizing the DMTOreactor design, and improving the operation efficiency.6.4.1. Fluidization of DMTO Catalyst. Although the

fluidization of Geldart type A particles such as FCC catalysthas been subject of research for decades, the knowledge onfluidization of DMTO catalyst is quite poor. It was found146

that the minimum fluidization velocity of DMTO catalystmeasured in the laboratory is lower than that predicted byAbrahamsen and Geldart’s empirical correlation,147 which iswidely used for FCC catalyst particles. Thus, a modifiedcorrelation for the minimum fluidization velocity of DMTOcatalyst was proposed by Zhao et al.146 An electrical capacitancetomography (ECT) method is being established to study theminimum bubbling velocity and the homogeneous fluidizationof DMTO catalyst.148

Another work is to develop internals for DMTO reactor tocontrol catalyst residence time distribution. As the catalystresidence time distribution is closely linked to the cokedistribution in the reactor, which in turn affects the selectivityto light olefins in DMTO process. By use of KCl particles astracers, we studied the influence of internals on catalystresidence time distribution in a pilot-scale cold flow fluidizedbed. On the basis of the results, an optimal internal wasdeveloped, and this will be implemented in the large scale coldflow fluidization facility in DICP. To further optimize DMTOfluidized bed reactor, a multifunctional cold-flow fluidizationfacility has been constructed in DICP. This facility includes aturbulent fluidized bed (inner diameter of 1.5 m), a bubblingfluidized bed (inner diameter of 1m), and a riser (innerdiameter of 30 cm, and length of 15 m).

6.4.2. Attrition of DMTO Catalyst. Some fundamentalresearch has been performed to study the attrition of theDMTO catalyst. It has been found that the DMTO catalyst isbetter than the typical FCC catalyst in the market in terms ofattrition resistance. Essentially, the attrition mechanisms indifferent operation regimes inside the DMTO reactor werestudied in the laboratory. Particularly, the focus is the attritionin the reaction conditions. Our results showed that the attritionmechanism of the DMTO catalyst at room temperature isdifferent from that at high temperature. At room temperature,both fragmentation and abrasion occur, although at hightemperature of 500 °C, the abrasion is dominant.149 Thismeans an optimal design of dust recovering system to capturesuperfine particles (less than 5 μm) might be essential.

6.4.3. DMTO Reactor Model. A reactor modeling approach,including hydrodynamics150 and kinetics,151,152 is beingestablished to simulate the DMTO fluidized bed reactor. Thehydrodynamics of the DMTO reactor in the demonstrationunit was simulated by use of an EMMS-based two-fluid CFDmodel (Figure 18). The results were compared with the

experimental data, and the solid fraction profile along the axialdirection show a good agreement.150 A lumped kinetic modelfor industrial DMTO catalyst was simultaneously developedbased on microscale reactor experiments.151,152 The kineticmodel can predicted the experimental data in the pilot-scaleDMTO reactor reasonably well. Via a collaboration with Prof.Jinghai Li’s group in Institute of Process Engineering, ChineseAcademy of Sciences, we are developing a comprehensivereactor model for industrial DMTO fluidized bed reactor.

6.5. DMTO Commercial Units. In May 2010, the world’sfirst DMTO commercial unit was constructed in Baotou, NorthChina by Shenhua group (Figure 19). The capacity of the unitis to produce 0.6 Mt polyethylene and polypropylene per year.

Figure 17. Reaction results in DMTO demonstration unit undercontinuous operation.

Table 7. Average Reaction Results in DMTO DemonstrationUnit

unit value

MeOH conversion wt % 99.42C2

= + C3= selectivity wt % 79.21

C2= + C3

= + C4= selectivity wt % 89.15

MeOH consumption t MeOH/t (C2= + C3

=) 2.96

Figure 18. CFD simulation of the DMTO fluidized bed reactor in thedemonstration unit.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1934

The reactor is a shallow turbulent fluidized bed with a diameterof 11 m and bed height of 3 m. Some engineering challengessuch as solid−gas distribution, catalyst mixing, and gasbreakthrough have to be solved in the design of this largeshallow fluidized bed. In August 2010, the unit was successfullystarted up. A method to start up the unit by use of the heatreleased from DMTO reactor was proposed. Compared toFCC process, this method can significantly reduce the time forstart up.153 Figure 20 shows the change of temperature in the

DMTO reactor during start up. As can be seen, the reactortemperature can rise by 100 °C in 60 min. When the reactorwas heated to 460 °C, the catalyst started to circulate betweenthe reactor and regenerator. It is worthy to note thatfundamental research on catalyst deactivation at low temper-ature (see Section 3.3) gave important information for thedevelopment of the start-up method. According to the research,there is a fast deactivation of the catalyst at lower temperaturerange (300−350 °C), and the catalyst should avoid thisdeactivation. Furthermore, circulation of catalyst should notstart until a high reaction temperature and high methanolconversion is achieved.Figure 21 is the methanol conversion and selectivity to light

olefins for the commercial unit, which are measured by theonline gas chromatography (GC). The methanol conversion is100%, and the selectivity is higher than 80%. The selectivity isbetter than the results in pilot-scale and demo-scale units,

which may be due to the good fluidization in commercialDMTO reactor. In March 2011, a 72 h performance test wascarried out for the Shenhua Baotou DMTO unit. On the basisof the test data, the methanol consumption is 2.96 tons per tonproduct of ethene and propene, which is the same as in thedemonstration unit. Up to now, this plant has produced morethan 2.1 million tons of polyolefins.

6.6. DMTO-II Process. After the success of DMTO process,DICP developed the second generation DMTO process(DMTO-II), in which the byproduct of C4

+ was separatedfrom the product stream and further converted to ethene andpropene in a second cracking fluidized bed reactor. Figure 22

shows the scheme of DMTO-II process. The main feature ofDMTO-II is that the MTO reaction and C4

+ conversionreaction share the same catalyst and could use one regeneratorfor catalyst regeneration. In 2009, the DMTO demonstrationunit was revamped by adding a second bubbling fluidized bedreactor for C4

+ cracking. From 2009 to 2010, the DMTO-IIprocess had been verified in the demonstration unit. The resultsobtained from the demonstration unit are listed in Table 8. Ascan be seen, the selectivity to ethene and propene increasedfrom 79.21% in DMTO process to 85.68% in the DMTO-IIprocess with 60 wt % of C4

+ recycling, and the methanolconsumption for producing one ton ethene and propenereduced from 2.96 to 2.67 tons. This apparently improves theeconomic profits for DMTO process. In October 26, 2010,

Figure 19. DMTO commercial unit in Shenhua Baotou.

Figure 20. Temperature increase in commercial DMTO reactorduring start up.

Figure 21. Methanol conversion and light olefins selectivity incommercial DMTO unit.

Figure 22. Scheme of DMTO-II process.

Table 8. Average Results of DMTO-II Demonstration Unit

unit value

MeOH conversion % 99.97C2

=+C3= selectivity wt % 85.68

MeOH consumption t MeOH/t (C2= + C3

=) 2.67C2

=/C3= ratio 0.7−1.1

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1935

DICP licensed the first DMTO-II unit (0.67 Mt/a polyethyleneand polypropylene product) to Pucheng Clean EnergyChemical Co. Ltd. In December 22, 2014, the first commercialDMTO-II unit was successfully started up.6.7. DMTO in China. Since 2006, DICP started to license

DMTO technology worldwide. Until December 31st, 2014, intotal 20 licenses were approved with a total capacity of 11 Mt/aethene and propene production. Six commercial DMTO unitsand one commercial DMTO-II unit have been put into streamwith olefins production capacity exceeding 4 Mt/a. More than10 units will be commissioned in near future according to theplan. It is expected that MTO will become an importantalternative to traditional light olefins’ production technologysuch as steam cracking of naphtha.

7. SUMMARY AND OUTLOOK

(1) MTO is an interesting and important reaction for bothfundamental research and industrial applications. Ourmechanism studies tend to prove that the HCPmechanism is generally effective for the reaction;however, the carbenium ions vary with the cavity size.Besides pore opening size of molecular sieves, the cavityseems to control selectivity as well. Although a reactionnetwork was proposed, there still remain many scientificchallenges, such as how the first C−C bond forms, whathappens in the induction period, what are the exactrelations among different reaction routes, how cokeforms, and how to control coking reaction in the reactionnetwork, among others. A better understanding of theMTO reaction is certainly critical for further improvingthe DMTO technology.

(2) Fundamental investigations on the SAPO molecular sievesynthesis offered strong support to the catalyst develop-ment. The successful development of the DMTO catalystsuggests that many essential issues need to be solved fora commercial applicable catalyst. The opportunities forfurther MTO catalyst innovation include the crystal-lization mechanism studies for precise control ofsynthesis, as well as the application of new catalyticmaterials such as meso-microporous zeolites that havebeen confirmed with longer lifetime for many reactions.To find a molecular sieve with suitable cavity sizerestricting coke formation is also a potential direction fordeveloping a better MTO catalyst.

(3) The development of DMTO technology is an integrationof many new analyses, including catalyst and engineeringaspects. The strategy to design DMTO catalyst withsimilar physical properties (density, particle size dis-tribution, etc.) to FCC catalyst, enabling the applicationof industrial FCC research, operation, and designexperience into the process development, is also criticalfor the rapid scale-up of DMTO reactor throughdemonstration test to commercialization.

(4) DMTO technology has been well developed anddemonstrated in China, which led to the successfulconstruction and commissioning of the world’s first coalto olefin plant. The application of DMTO technology isexpected to change the strategy and framework for thesupply of olefins in China. It can also play an importantrole in other countries abundant in coal or natural gas.

■ ASSOCIATED CONTENT*S Supporting InformationThe following file is available free of charge on the ACSPublications website at DOI: 10.1021/acscatal.5b00007.

Detailed comparison between DMTO and FCC process(PDF)

■ AUTHOR INFORMATIONCorresponding Author*E-mail: liuzm@dicp.ac.cn. Tel.: 0086-411-84379998.NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSWe are very grateful to the support from Chinese Academy ofSciences (CAS), Chinese National Development and Reform-ing Committee (NDRC), Chinese Ministry of Science andTechnology, National Natural Science Foundation of China(NSFC), Government of Shanxi Province, China Petroleumand Chemical Industry Federation, Shenhua Group Corpo-ration Limited. We thank our partners SINOPEC LuoyangPetrochemical Engineering Co. and SYN Energy TechniqueCo., and all colleagues involved in the DMTO processdevelopment.

■ REFERENCES(1) Chang, C. D.; Silvestri, A. J. J. Catal. 1977, 47, 249−259.(2) Stocker, M. Microporous Mesoporous Mater. 1999, 29, 3−48.(3) Wang, W.; Hunger, M. Acc. Chem. Res. 2008, 41, 895−904.(4) Haw, J. F.; Song, W. G.; Marcus, D. M.; Nicholas, J. B. Acc. Chem.Res. 2003, 36, 317−326.(5) Arstad, B.; Kolboe, S. J. Am. Chem. Soc. 2001, 123, 8137−8138.(6) Chen, D.; Moljord, K.; Holmen, A. Microporous MesoporousMater. 2012, 164, 239−250.(7) Svelle, S.; Joensen, F.; Nerlov, J.; Olsbye, U.; Lillerud, K.-P.;Kolboe, S.; Bjorgen, M. J. Am. Chem. Soc. 2006, 128, 14770−14771.(8) Bjorgen, M.; Akyalcin, S.; Olsbye, U.; Benard, S.; Kolboe, S.;Svelle, S. J. Catal. 2010, 275, 170−180.(9) Bjorgen, M.; Olsbye, U.; Petersen, D.; Kolboe, S. J. Catal. 2004,221, 1−10.(10) Dai, W.; Wang, X.; Wu, G.; Guan, N.; Hunger, M.; Li, L. ACSCatal. 2011, 1, 292−299.(11) Cao, G.; Matu J. S. U.S. Patent 6,680,278, Jan. 20, 2004.(12) Chang, Y.; Vaughn, S. N.; Martens, L. R. M. U.S. Patent7,271,123, Sep. 18, 2007.(13) Mertens, M.; Strohmaier, K. G. U.S. Patent 6,773,688, Aug. 10,2004.(14) Liu, G.; Tian, P.; Liu, Z. HuaXueJinZhan 2010, 22, 1531−1537.(15) Wilson, S.; Barger, P. Microporous Mesoporous Mater. 1999, 29,117−126.(16) Vora, B. V.; Marker, T. L.; Barger, P. T.; Nilsen, H. R.; Kvisle, S.;Fuglerud, T. Stud. Surf. Sci. Catal. 1997, 107, 87−98.(17) Chen, J. Q.; Bozzano, A.; Glover, B.; Fuglerud, T.; Kvisle, S.Catal. Today 2005, 106, 103−107.(18) Coute, N. P.; Kuechler, K. H.; Chisholm, P. N.; Vaughn, S. N.;Lattner, J. R.; Kuechler Sr., W. L. U.S. Patent 6,673,978, Jan. 6, 2004.(19) Koempel, H.; Liebner, W. Stud. Surf. Sci. Catal. 2007, 167, 261−267.(20) Olsbye, U.; Svelle, S.; Bjorgen, M.; Beato, P.; Janssens, T. V. W.;Joensen, F.; Bordiga, S.; Lillerud, K. P. Angew. Chem., Int. Ed. 2012, 51,5810−5831.(21) Liu, Z.; Liang, J. Curr. Opin. Solid State Mater. Sci. 1999, 4, 80−84.(22) Liu, Z.; Sun, C.; Wang, G.; Wang, Q.; Cai, G. Fuel Process.Technol. 2000, 62, 161−172.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1936

(23) Chen, N. Y.; Reagan, W. J. J. Catal. 1979, 59, 123−129.(24) Ono, Y.; Mori, T. J. Chem. Soc. Fara. Trans. I 1981, 77, 2209−2221.(25) Chang, C. D. Catal. Rev. 1983, 25, 1−118.(26) Swabb, E. A.; Gates, B. C. Ind. Eng. Chem. Fundam. 1972, 11,540−545.(27) Hutchings, G. J.; Gottschalk, F.; Hall, M. V. M; Hunter, R. J.Chem. Soc. Fara. Trans. I 1987, 83, 571−583.(28) Olah, G. A. Pure Appl. Chem. 1981, 53, 201−207.(29) Kagi, D. J. Catal. 1981, 69, 242−243.(30) Clarke, J. K. A.; Darcy, R.; Hegarty, B. F.; Odonoghue, E.;Amirebrahimi, V.; Rooney, J. J. J. Chem. Soc., Chem. Commun. 1986,425−426.(31) Zatorski, W.; Kryzanowski, S. Acta Phys. Chem. 1977, 24, 347−352.(32) Dessau, R. M. J. Catal. 1986, 99, 111−116.(33) Dessau, R. M.; Lapierre, R. B. J. Catal. 1982, 78, 136−141.(34) Mole, T.; Bett, G.; Seddon, D. J. Catal. 1983, 84, 435−445.(35) Mole, T.; Whiteside, J. A.; Seddon, D. J. Catal. 1983, 82, 261−266.(36) Dahl, I. M.; Kolboe, S. Catal. Lett. 1993, 20, 329−336.(37) Dahl, I. M.; Kolboe, S. J. Catal. 1994, 149, 458−464.(38) Dahl, I. M.; Kolboe, S. J. Catal. 1996, 161, 304−309.(39) Hunger, M. Catal. Rev. 1997, 39, 345−393.(40) Wang, W.; Jiang, Y. J.; Hunger, M. Catal. Today 2006, 113, 102.(41) Xu, T.; White, J. L. U.S. Patent 6,743,747, Jun. 1, 2004.(42) Xu, T.; White, J. L. U.S. patent 6,734,330, May 11, 2004.(43) Arstad, B.; Kolboe, S. Catal. Lett. 2001, 71, 209−212.(44) Fu, H.; Song, W. G.; Haw, J. F. Catal. Lett. 2001, 76, 89−94.(45) Haw, J. F.; Marcus, D. M. Top. Catal. 2005, 34, 41−48.(46) Song, W. G.; Fu, H.; Haw, J. F. J. Am. Chem. Soc. 2001, 123,4749−4754.(47) Song, W. G.; Fu, H.; Haw, J. F. J. Phys. Chem. B 2001, 105,12839−12843.(48) Song, W. G.; Haw, J. F.; Nicholas, J. B.; Heneghan, C. S. J. Am.Chem. Soc. 2000, 122, 10726−10727.(49) Haw, J. F.; Nicholas, J. B.; Song, W. G.; Deng, F.; Wang, Z. K.;Xu, T.; Heneghan, C. S. J. Am. Chem. Soc. 2000, 122, 4763−4774.(50) Song, W. G.; Nicholas, J. B.; Sassi, A.; Haw, J. F. Catal. Lett.2002, 81, 49−53.(51) Xu, T.; Barich, D. H.; Goguen, P. W.; Song, W. G.; Wang, Z. K.;Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998, 120, 4025−4026.(52) Xu, T.; Haw, J. F. J. Am. Chem. Soc. 1994, 116, 10188−10195.(53) Wang, C.; Chu, Y. Y.; Zheng, A. M.; Xu, J.; Wang, Q.; Gao, P.;Qi, G. D.; Gong, Y. J.; Deng, F. Chem.- Eur. J. 2014, 20, 12432−12443.(54) Sassi, A.; Wildman, M. A.; Ahn, H. J.; Prasad, P.; Nicholas, J. B.;Haw, J. F. J. Phys. Chem. B 2002, 106, 2294−2303.(55) Sassi, A.; Wildman, M. A.; Haw, J. F. J. Phys. Chem. B 2002, 106,8768−8773.(56) Arstad, B.; Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 2004,126, 2991−3001.(57) Bjorgen, M.; Bonino, F.; Kolboe, S.; Lillerud, K. P.; Zecchina,A.; Bordiga, S. J. Am. Chem. Soc. 2003, 125, 15863−15868.(58) McCann, D. M.; Lesthaeghe, D.; Kletnieks, P. W.; Guenther, D.R.; Hayman, M. J.; Van Speybroeck, V.; Waroquier, M.; Haw, J. F.Angew. Chem., Int. Ed. 2008, 47, 5179−5182.(59) Nicholas, J. B.; Haw, J. F. J. Am. Chem. Soc. 1998, 120, 11804−11805.(60) Bjorgen, M.; Svelle, S.; Joensen, F.; Nerlov, J.; Kolboe, S.;Bonino, F.; Palumbo, L.; Bordiga, S.; Olsbye, U. J. Catal. 2007, 249,195−207.(61) Svelle, S.; Olsbye, U.; Joensen, F.; Bjorgen, M. J. Phys. Chem. C2007, 111, 17981−17984.(62) Guisnet, M.; Costa, L.; Ribeiro, F. R. J. Mol. Catal. A: Chem.2009, 305, 69−83.(63) Liu, Z.; Qi, Y. Bull. Chin. Acad. Sci. 2006, 21, 406−408.(64) Li, J. Ph.D. Dissertation, Dalian Institute of Chemical Physics,2008.

(65) Li, J.; Qi, Y.; Liu, Z.; Liu, G.; Chang, F. Chin. J. Catal. 2008, 29,660−664.(66) Li, J.; Qi, Y.; Liu, Z.; Liu, G.; Zhang, D. Catal. Lett. 2008, 121,303−310.(67) Li, J.; Qi, Y.; Xu, L.; Liu, G.; Meng, S.; Li, B.; Li, M.; Liu, Z.Catal. Commun. 2008, 9, 2515−2519.(68) Li, J.; Wei, Y.; Liu, G.; Qi, Y.; Tian, P.; Li, B.; He, Y.; Liu, Z.Catal. Today 2011, 171, 221−228.(69) Wang, C.; Wang, Y.; Xie, Z.; Liu, Z. J. Phys. Chem. C 2009, 113,4584−4591.(70) Su, X.; Tian, P.; Fan, D.; Xia, Q.; Yang, Y.; Xu, S.; Zhang, L.;Zhang, Y.; Wang, D.; Liu, Z. ChemSusChem 2013, 6, 911−918.(71) Su, X.; Tian, P.; Li, J.; Zhang, Y.; Meng, S.; He, Y.; Fan, D.; Liu,Z. Microporous Mesoporous Mater. 2011, 144, 113−119.(72) Tian, P.; Su, X.; Wang, Y.; Xia, Q.; Zhang, Y.; Fan, D.; Meng, S.;Liu, Z. Chem. Mater. 2011, 23, 1406−1413.(73) Li, J.; Wei, Y.; Chen, J.; Tian, P.; Su, X.; Xu, S.; Qi, Y.; Wang, Q.;Zhou, Y.; He, Y.; Liu, Z. J. Am. Chem. Soc. 2012, 134, 836−839.(74) Li, J.; Wei, Y.; Xu, S.; Tian, P.; Chen, J.; Liu, Z. Catal. Today2014, 226, 47−51.(75) Xu, S.; Zheng, A.; Wei, Y.; Chen, J.; Li, J.; Chu, Y.; Zhang, M.;Wang, Q.; Zhou, Y.; Wang, J.; Deng, F.; Liu, Z. Angew. Chem., Int. Ed.2013, 52, 11564−11568.(76) Chen, J.; Li, J.; Wei, Y.; Yuan, C.; Li, B.; Xu, S.; Zhou, Y.; Wang,J.; Zhang, M.; Liu, Z. Catal. Commun. 2014, 46, 36−40.(77) Li, J.; Wei, Y.; Chen, J.; Xu, S.; Tian, P.; Yang, X.; Li, B.; Wang,J.; Liu, Z. M. ACS Catal. 2014, 5, 661−665.(78) Yuan, C.; Wei, Y.; Li, J.; Xu, S.; Chen, J.; Zhou, Y.; Wang, Q.;Xu, L.; Liu, Z. Chin. J. Catal. 2012, 33, 367−374.(79) Wei, Y.; Li, J.; Yuan, C.; Xu, S.; Zhou, Y.; Chen, J.; Wang, Q.;Zhang, Q.; Liu, Z. Chem. Commun. 2012, 48, 3082−3084.(80) Wei, Y.; Yuan, C.; Li, J.; Xu, S.; Zhou, Y.; Chen, J.; Wang, Q.;Xu, L.; Qi, Y.; Zhang, Q.; Liu, Z. ChemSusChem 2012, 5, 906−912.(81) Yuan, C.; Wei, Y.; Xu, L.; Li, J.; Xu, S.; Zhou, Y.; Chen, J.; Wang,Q.; Liu, Z. Chin. J. Catal. 2012, 33, 768−770.(82) Chen, J.; Li, J.; Yuan, C.; Xu, S.; Wei, Y.; Wang, Q.; Zhou, Y.;Wang, J.; Zhang, M.; He, Y.; Xu, S.; Liu, Z. Catal. Sci. Technol. 2014, 4,3268−3277.(83) Brent, M. L.; Celeste, A. M.; Robert, L. P.; Richard, T. G.;Thomas, R. C.; Edith, M. F. U.S. Patent 4,440,871, Apr. 3, 1984.(84) Brent, M. L.; Celeste, A. M.; R. Lyle, P.; Richard, T. G.; Thomas,R. C.; Edith, M. F. J. Am. Chem. Soc. 1984, 106, 6092−6093.(85) Liang, J.; Li, H. Y.; Zhao, S.; Guo, W. G.; Wang, R. H.; Ying, M.L. Appl. Catal. 1990, 64, 31−40.(86) Chen, D.; Moljord, K.; Fuglerud, T.; Holmen, A. MicroporousMesoporous Mater. 1999, 29, 191−203.(87) Vomscheid, R.; Briend, M.; Peltre, M. J.; Man, P. P.;Barthomeuf, D. J. Phys. Chem. 1994, 98, 9614−9618.(88) Briend, M.; Vomscheid, R.; Peltre, M. J.; Man, P. P.;Barthomeuf, D. J. Phys. Chem. 1995, 99, 8270−8276.(89) Sastre, G.; Lewis, D. W.; Catlow, C. R. A. J. Phys. Chem. 1996,100, 6722−6730.(90) Bleken, F.; Bjørgen, M.; Palumbo, L.; Bordiga, S.; Svelle, S.;Lillerud, K.-P.; Olsbye, U. Top. Catal. 2009, 52, 218−228.(91) Xi, D.; Sun, Q.; Xu, J.; Cho, M.; Cho, H. S.; Asahina, S.; Li, Y.;Deng, F.; Terasaki, O.; Yu, J. J. Mater. Chem. A 2014, 2, 17994−18004.(92) Dargahi, M.; Kazemian, H.; Soltanieh, M.; Rohani, S.;Hosseinpour, M. Particuology 2011, 9, 452−457.(93) Vistad, Ø. B.; Akporiaye, D. E.; Taulelle, F.; Lillerud, K. P.Chem. Mater. 2003, 15, 1639−1649.(94) Martins, G. A. V.; Berlier, G.; Coluccia, S.; Pastore, H. O.;Superti, G. B.; Gatti, G.; Marchese, L. J. Phys. Chem. C 2006, 111,330−339.(95) Wu, X.; Abraha, M. G.; Anthony, R. G. Appl. Catal., A 2007,260, 63−69.(96) Nishiyama, N.; Kawaguchi, M.; Hirota, Y.; Van Vu, D.; Egashira,Y.; Ueyama, K. Appl. Catal., A 2009, 362, 193−199.(97) Buchholz, A.; Wang, W.; Arnold, A.; Xu, M.; Hunger, M.Microporous Mesoporous Mater. 2003, 57, 157−168.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1937

(98) Hirota, Y.; Murata, K.; Miyamoto, M.; Egashira, Y.; Nishiyama,N. Catal. Lett. 2010, 140, 22−26.(99) Zhang, L.; Bates, J.; Chen, D.; Nie, H. Y.; Huang, Y. J. Phys.Chem. C 2011, 115, 22309−22319.(100) Li, Z.; Martinez-Triguero, J.; Concepcion, P.; Yu, J.; Corma, A.Phys. Chem. Chem. Phys. 2013, 15, 14670−14680.(101) Li, H.; Liang, J.; Wang, R.; Liu, Z.; Zhao, S. Petrochem. Technol.1987, 16, 340−346.(102) Liu, Z.; Huang, X.; He, C.; Yang, Y.; Yang, L.; Cai, G. Chin. J.Catal. 1996, 17, 540−543.(103) Liu, Z.; Sun, C.; Xu, L.; Yang, L.; Tian, P.; Tan, J. CN Patent1131845C, Dec. 24, 2003.(104) Liu, Z.; Xu, L.; Tian, P.; Yang, Y.; Yang, L.; He, C.; Lv, Z.;Meng, S.; Wang, G.; Sun, X.; Yang, H.; Wang, X.; Yuan, C.; Li, M.;Wei, Y.; Qi, Y.; Zhu, S.; Zhang, J.; Lu, X.; Wang, H.; Xie, P.; Li, B. CNPatent 100584758C, Jan. 27, 2010.(105) Liu, Z.; Cai, G.; He, C.; Yang, L.; Wang, Z.; Luo, J.; Chang, Y.;Shi, R.; Jiang, Z.; Sun, C. CN Patent 1037334C, Feb. 11, 1998.(106) Tan, J. Mater Dissertation, Dalian Institute of ChemicalPhysics, 1999.(107) Du, A. Mater Dissertation, Dalian Institute of ChemicalPhysics, 2004.(108) Zhang, D. Ph.D. Dissertation, Dalian Institute of ChemicalPhysics, 2007.(109) Liu, G. Ph.D. Dissertation, Dalian Institute of ChemicalPhysics, 2008.(110) Tan, J.; Liu, Z. M.; Bao, X. H.; Liu, X. C.; Han, X. W.; He, C.Q.; Zhai, R. S. Microporous Mesoporous Mater. 2002, 53, 97−108.(111) Tian, P.; Li, B.; Xu, S.; Su, X.; Wang, D.; Zhang, L.; Fan, D.;Qi, Y.; Liu, Z. J. Phys. Chem. C 2013, 117, 4048−4056.(112) Li, B.; Tian, P.; Qi, Y.; Zhang, L.; Xu, S.; Su, X.; Fan, D.; Liu, Z.Chin. J. Catal. 2013, 34, 593−603.(113) Liu, G. Y.; Tian, P.; Zhang, Y.; Li, J. Z.; Xu, L.; Meng, S. H.;Liu, Z. M. Microporous Mesoporous Mater. 2008, 111, 143−149.(114) Tian, P.; Liu, Z. M.; Xu, l.; Sun, C. Stud. Surf. Sci. Catal. 2001,135, 248−248.(115) Xu, L.; Liu, Z. M.; Tian, P.; Wei, Y.; Sun, C.; Li, S. Stud. Surf.Sci. Catal. 2001, 135, 247−247.(116) Xu, L.; Liu, Z. M.; Du, A. P.; Wei, Y. X.; Sun, Z. G. Stud. Surf.Sci. Catal. 2004, 147, 445−450.(117) Liu, Z.; Xu, L.; Sun, C.; Huang, T.; Tian, P.; Yang, L.; Tan, J.EP Patent 1,142,833, July 10, 2009.(118) Liu, G. Y.; Tian, P.; Li, J. Z.; Zhang, D. Z.; Zhou, F.; Liu, Z. M.Microporous Mesoporous Mater. 2008, 114, 416−423.(119) Barthomeuf, D. Zeolites 1994, 14, 394−401.(120) Mores, D.; Stavitski, E.; Kox, M. H.; Kornatowski, J.; Olsbye,U.; Weckhuysen, B. M. Chem. − Eur. J. 2008, 14, 11320−11327.(121) Liu, G.; Tian, P.; Zhang, Y.; Li, J.; Xu, L.; Meng, S.; Liu, Z.Microporous Mesoporous Mater. 2008, 111, 143−149.(122) Izadbakhsh, A.; Farhadi, F.; Khorasheh, F.; Sahebdelfar, S.;Asadi, M.; Yan, Z. F. Microporous Mesoporous Mater. 2009, 126, 1−7.(123) Xu, L.; Du, A.; Wei, Y.; Meng, S.; He, Y.; wang, Y.; Yu, Z.;Zhang, X.; Liu, Z. Chin. J. Catal. 2008, 29, 727−732.(124) Fan, D.; Tian, P.; Xu, S.; Xia, Q.; Su, X.; Zhang, L.; Zhang, Y.;He, Y.; Liu, Z. J. Mater. Chem. 2012, 22, 6568−6574.(125) Fan, D.; Tian, P.; Su, X.; Yuan, Y.; Wang, D.; Wang, C.; Yang,M.; Wang, L.; Xu, S.; Liu, Z. J. Mater. Chem. A 2013, 1, 14206−14213.(126) Wang, D.; Tian, P.; Yang, M.; Xu, S.; Fan, D.; Su, X.; Yang, Y.;Wang, C.; Liu, Z. Microporous Mesoporous Mater. 2014, 194, 8−14.(127) van Heyden, H.; Mintova, S.; Bein, T. Chem. Mater. 2008, 20,2956−2963.(128) Zhu, J.; Cui, Y.; Wang, Y.; Wei, F. Chem. Commun. 2009,3282−3284.(129) Yang, H.; Liu, Z.; Gao, H.; Xie, Z. J. Mater. Chem. 2010, 20,3227−3231.(130) Schmidt, F.; Paasch, S.; Brunner, E.; Kaskel, S. MicroporousMesoporous Mater. 2012, 164, 214−221.(131) Sun, Q.; Wang, N.; Xi, D.; Yang, M.; Yu, J. Chem. Commun.2014, 50, 6502−6505.

(132) Hirota, Y.; Murata, K.; Tanaka, S.; Nishiyama, N.; Egashira, Y.;Ueyama, K. Mater. Chem. Phys. 2010, 123, 507−509.(133) Yang, M.; Tian, P.; Wang, C.; Yuan, Y.; Yang, Y.; Xu, S.; He,Y.; Liu, Z. Chem. Commun. 2014, 50, 1845−1847.(134) Wang, C.; Yang, M.; Tian, P.; Xu, S.; Yang, Y.; Wang, D.; Yuan,Y.; Liu, Z. J. Mater. Chem. A 2015, accepted.(135) Liu, Z.; Tian, P.; Xu, L.; Yang, L.; Lv, Z.; Qi, Y.; He, C.; Wei,Y.; Zhang, J.; Meng, S.; Li, M.; Yuan, C.; Wang, X.; Yang, Y.; Lu, X.;Zhu, S.; Xie, P.; Sun, X.; Yang, H.; Wang, H.; Li, B. CN Patent101122145C, Oct. 20, 2010.(136) Tian, P.; Liu, Z.; Xu, L.; Yang, L.; Qi, Y.; Wang, X.; Yuan, C.;Lv, Z.; Meng, S. CN Patent 101121146B, Sep. 14, 2011.(137) Tian, P.; Liu, Z.; Xu, L.; Sang, S.; He, C.; Yang, L.; Yuan, C.CN Patent 101157057B, Dec. 1, 2010.(138) Miller, L. W. U.S. Patent 6,166,282, Dec. 26, 2000.(139) Geldart, D. Powder Technol. 1973, 7, 285−292.(140) Du, B.; Fan, L.; Wei, F.; Warsito, W. AIChE J. 2002, 48, 1896−1909.(141) Knowlton, T. M.; Karri, S. B. R.; Issangya, A. Powder Technol.2005, 150, 72−77.(142) Matsen, J. M. Powder Technol. 1996, 88, 237−244.(143) Rudisuli, M.; Schildhauera, T. J.; Biollaz, S. M. A.; van Ommen,J. R. Powder Technol. 2012, 217, 21−38.(144) Gauthier, T. A. Int. J. Chem. Reactor Eng. 2009, 7, A22.(145) Werther, J. AIChE symposium series 1974, 70, 53−62.(146) Zhao, Y.; Hao, J.; Ye, M.; Liu, Z. Proceedings of the 11thInternational Conference on Fluidized Bed Technology (CFB-11),Beijing, Chian, May 14−17, 2014; Li, J. H., Eds; Chemical IndustryPress: Beijing, 2014.(147) Abrahamsen, A. R.; Geldart, D. Powder Technol. 1980, 26, 35−46.(148) Luo, Q.; Zhao, Y.; Ye, M.; Liu, Z. CIESC Journal 2014, 65,2504−2512.(149) Ying, L.; Ye, M.; Cheng, Y.; Li, X.; Liu, Z. Proceedings of the14th International Conference on Fluidization, Noordwijkerhout,Netherlands, May 26-31, 2013; Kuipers, J. A. M.; Mudde, R. F.; vanOmmen, J. R., Deen, N. G., Eds; ECI Symposium Series, 2013.(150) Hao, J.; Zhao, Y.; Ye, M.; Liu, Z. Adv. Powder Technol. 2015, inrevision.(151) Zhao, Y.; Li, H.; Ye, M.; Liu, Z. Ind. Eng. Chem. Res. 2013, 52,11354−11364.(152) Ying, L.; Ye, M.; Cheng, Y.; Li, H.; Liu, Z. Chem. Eng. Res. Des.Submitted for publication, 2015.(153) Liu, Z.; Lv, Z.; He, C.; Liu, Y.; Qi, Y.; Min, X.; Wang, G.;Wang, X.; Zhang, J. CN Patent 101130466B, May 4, 2011.

ACS Catalysis Perspective

DOI: 10.1021/acscatal.5b00007ACS Catal. 2015, 5, 1922−1938

1938